INTRODUCTION AND AIMS: To describe the characteristics of non-fatal medication-related ambulance attendances in Melbourne. DESIGN AND METHODS: A retrospective analysis of 16 705 patient care records completed by ambulance paramedics in Melbourne where medications had a causal role in the attendance. RESULTS: A single medication only was implicated in 11 765 cases (70% of the total). Of these, 85% involved one of six types of medication: benzodiazepines (52%), paracetamol (15%), selective serotonin re-uptake inhibitors (6.5%), combination paracetamol and opioids (4%), phenothiazines (3.4%) and tricyclic antidepressants (TCA) (3.7%). Cases involving benzodiazepines were significantly (P < 0.001) older (Average = 37 years) than those involving paracetamol (Average = 30 years). Thirty-four per cent of cases involved concurrent alcohol use, and this varied according to drug type (paracetamol 26%, benzodiazepines 40%, selective serotonin re-uptake inhibitors 35%, paracetamol and opioids 35%). An abnormal Glasgow Coma Scale score was found in 19% of cases, again varying according to drug type (paracetamol 10%, TCA 39%, benzodiazepines 21%, paracetamol and opioids 17%, phenothiazines 15%). Ten per cent of cases were not transported to hospital ranging from 3% for TCA to 13% for benzodiazepines. DISCUSSION AND CONCLUSIONS: The majority of non-fatal medication events attended by ambulance paramedics involve one of six substances. Benzodiazepines were most commonly implicated and, as management may require only simple supportive treatment, significant numbers are not transported to hospital. The unique clinical population is identified in this study and the ongoing medical and psychiatric treatment of these patients not transported to hospital in the study period needs to be considered.